OMITTED

What the news leaves out.

← Omitted front page

AI Risks Nearly Sparked US-China Conflict

6 sources · updated 2026-10-01
Left 67% Center 0% Right 33%
2 left · 0 center · 1 right

What happened

Palisade Research, a nonprofit that studies AI capabilities and motivations, published video interviews on frominside.ai with current and former researchers from OpenAI, Google DeepMind, and Anthropic who warned that superintelligent AI could cause human extinction. In those interviews, former OpenAI and Google DeepMind employee Geoffrey Irving said the chance of human extinction is "about a coin flip," Google DeepMind research scientist Neel Nanda said there is "at least a 10 percent chance that it causes human extinction," and former OpenAI researcher Daniel Kokotajlo said superintelligent AI "would basically be god-like powerful" and that "unfortunately we don't know how to control them at all." Anthropic engineer Jacob Coxon resigned, saying OpenAI and Anthropic are "racing straight towards self-improving superintelligence and gambling with our lives"; California Governor Gavin Newsom issued a mandate to create an "AI kill-switch," New York Governor Kathy Hochul began exploring similar measures, and Senator Bernie Sanders and Representative Greg Casar introduced legislation to "ban artificial superintelligence." A paper coauthored by GovAI research fellow Alan Chan with Anthropic cofounder Jack Clark, OpenAI chief scientist Jakub Pachocki, Microsoft chief scientific officer Eric Horvitz, and Geoffrey Hinton called on policymakers to require insight into labs' automation of AI research, which Chan termed an "intelligence explosion"; leaders of Anthropic, Google, OpenAI, xAI, and Nvidia signed a White House accord to "partner with an independent external auditor or evaluator," which President Trump described as "morally" binding. A CNN report published on September 18 described U.S. military plans to intercept a Chinese vessel after an intelligence report generated with a chatbot claimed the ship was transporting nuclear-weapon components; according to that report, officials stopped the operation after further investigation found the claim was erroneous.
Sourcing check · Fully on the record · 3 articles

1 of 1 article cites a named source for the story. It was on the record from the first report.

What these terms mean and how they're counted
On the record · names a source 1 of 1
On the record means the article cites a source by name for the story's main claim: a person, or an institution speaking officially, such as a company statement or a named agency's spokesperson. The opposite is an unnamed source. The count is how many of the counted articles cite at least one such source; the names are everyone named across them, each once however the articles spell it. Here: Alan Chan and Anthropic, Google, OpenAI, SpaceXAI, and Nvidia
Primary evidence · cites a document or data 0 of 1
Primary evidence means the article cites a record you could check yourself: a court filing or ruling, a dataset or report, a recording or transcript, an official statement or press release, market data. A quote is not a document. The count is how many articles cite one. Here: None cited.
Credits another outlet 0 of 1
The article says the story comes from another outlet's reporting (“first reported by Reuters”). Those articles relay the story; they are not independent confirmation. Here: Each article presents its own reporting.
Unnamed sources 0 of 1
The main claim is attributed to people who are not identified: “people familiar with the matter”, “officials said”. The count is how many articles do this.
How it developed
When the first report landed and whether it named anyone; when the first named source and the first document appeared. Here: First report Oct 1, 10:09 UTC by Business Insider, naming its source.
Not counted 2
Opinion, analysis and video pieces, and articles too short to assess. Every count above is out of the counted articles. An article republished under a new headline is one article.
This is about who is cited, not whether the story is true. Each source below says what it names. How we count
Omitted — what each side leaves out

Unpacked

Right-leaning coverage is simply less complete: it never reports the researchers' warnings, so readers on that side miss the claims that made this a story. Left-leaning coverage at the Verge carries interviews in which lab insiders call superintelligence "exactly as dangerous as it sounds." A DeepMind scientist in those interviews assigned at least a 10 percent chance of human extinction. National Review instead argues that giving control of AI to tech leaders and politicians is more dangerous than letting the public manage the risks. Left-leaning coverage leads with current and former staff at OpenAI, Google, and Anthropic. Right-leaning coverage leads with elected officials, pairing its case with a photograph of Senator Bernie Sanders. The coverage never asks what evidence produced the extinction percentages those researchers cite.
Bottom line

National Review never prints a lab researcher's extinction number. The Verge's 10 percent figure from a DeepMind scientist stays on the left.

The receipts — 1 fact one side's coverage lacks
  • A paper co-written with Jack Clark, Jakub Pachocki, Eric Horvitz, and Geoffrey Hinton urges required embedded auditors of automated AI development contextin Business Insider · absent from right-leaning coverage
The Left View
Left-leaning coverage both amplified the researchers' extinction warnings and treated those warnings as a marketing diversion from present-day harms. The Verge presented the Palisade videos as industry insiders echoing "doomer" claims, highlighting Kokotajlo's statement that the situation "is exactly as dangerous as it sounds and must not be allowed to happen," and researchers' reasons for remaining at the labs—including Google's Mary Phuong saying "I think you absolutely should be suspicious of what I'm saying because I am being paid by the lab" and Nanda saying that if he did not believe his work was "directly reducing these existential risks, I would quit." In the Guardian, Timnit Gebru and Emily M. Bender argued that while kill-switch mandates and an ASI ban captured attention, the undercovered CNN military episode showed that "government reliance on error-prone systems" marketed as near-superintelligent nearly produced "armageddon"; they described today's models as "stochastic parrots" that generate "plausible-looking text," wrote that "the harmful, and well-documented, impacts of these companies' products have arisen because people believe that the products are superintelligent, not because they actually are," and said that describing the products as "beings with agency allows these executives to evade accountability." Business Insider's account of Chan's paper emphasized the prospect of "Mythos-level moments, not every year, but every month," the authors' warning that "full automation within this timeframe should be taken seriously," and Chan's claim that policymakers face "a very opportune moment" whose impact "may diminish over time."
The Right View
National Review framed the controversy as "The Dangers of Misaligned Human Intelligence," arguing that "Giving control of AI to tech leaders and politicians is more dangerous than letting all of us manage its risks." The piece was illustrated with an image of Senator Bernie Sanders, a sponsor of the legislation to ban artificial superintelligence.
Our Take (balanced)
The left's strongest argument from inside the labs is that people building frontier systems assign extinction probabilities from "at least a 10 percent chance" to "about a coin flip" and describe superintelligence as "god-like" technology "we don't know how to control," a concern echoed in a paper whose coauthors include senior figures at Anthropic, OpenAI, and Microsoft treating automated AI research as a near-term possibility that warrants outside auditors while a policy window remains. A second left argument, from Gebru and Bender, is empirically anchored in the CNN-reported military chatbot incident and in documented civilian harms, holding that "misplaced faith in unreliable automated software"—not hypothetical rogue superintelligence—is the demonstrated danger, and that agency talk lets vendors "evade accountability." The right's strongest argument is that the more serious misalignment is human: concentrating authority over AI in tech leaders and the politicians writing bans and kill-switches, rather than leaving risk management broadly distributed. The central unresolved tension is whether the primary threat is machines that may escape anyone's control, or the concentration of control over those machines in a small set of executives and officials.

6 sources

More stories

The week's bottom lines, in your inbox

One email a week: the five stories that mattered and what they actually mean. Free.